US8346509B2 - Context switch sampling - Google Patents

Context switch sampling Download PDF

Info

Publication number
US8346509B2
US8346509B2 US12/420,815 US42081509A US8346509B2 US 8346509 B2 US8346509 B2 US 8346509B2 US 42081509 A US42081509 A US 42081509A US 8346509 B2 US8346509 B2 US 8346509B2
Authority
US
United States
Prior art keywords
hardware event
hardware
event
counter
event counter
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related, expires
Application number
US12/420,815
Other languages
English (en)
Other versions
US20100262870A1 (en
Inventor
Robert Davies
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Intel Corp
Original Assignee
Intel Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Intel Corp filed Critical Intel Corp
Priority to US12/420,815 priority Critical patent/US8346509B2/en
Assigned to INTEL CORPORATION reassignment INTEL CORPORATION ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: DAVIES, ROBERT
Priority to JP2010088905A priority patent/JP5257816B2/ja
Priority to EP10250736.5A priority patent/EP2239664B1/de
Priority to CN201010207270.1A priority patent/CN101859268B/zh
Publication of US20100262870A1 publication Critical patent/US20100262870A1/en
Application granted granted Critical
Publication of US8346509B2 publication Critical patent/US8346509B2/en
Expired - Fee Related legal-status Critical Current
Adjusted expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
    • G06F11/3466—Performance evaluation by tracing or monitoring
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
    • G06F11/3017—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system is implementing multitasking
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3089—Monitoring arrangements determined by the means or processing involved in sensing the monitored data, e.g. interfaces, connectors, sensors, probes, agents
    • G06F11/3093—Configuration details thereof, e.g. installation, enabling, spatial arrangement of the probes
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2201/00—Indexing scheme relating to error detection, to error correction, and to monitoring
    • G06F2201/86—Event-based monitoring
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2201/00—Indexing scheme relating to error detection, to error correction, and to monitoring
    • G06F2201/88—Monitoring involving counting

Definitions

  • This disclosure relates generally to the field of microprocessors.
  • the disclosure relates to context switch sampling of process or thread events based on hardware event triggers in a processor.
  • monitoring performance metrics may be complicated.
  • Techniques that have been used in the past such as time-based sampling or event-based sampling employ a consistent regular grid of measurement to outline and characterize the behavior of applications whose activity may at times be anything but regular.
  • Previous attempts to monitor activity within the context of a particular process may have required specially instrumented versions of the operating system. These techniques may also have the side effect of monitoring the special instrumentation as well as the desired performance metrics in the context of the particular process of interest. Thus results of previous techniques may have been contaminated by activity from other processes, threads or operating system instrumentation.
  • FIG. 1 illustrates one embodiment of a multiprocessing system for performing context switch sampling.
  • FIG. 2 illustrates a flow diagram for one embodiment of a process for performing context switch sampling.
  • FIG. 3 illustrates a flow diagram for an alternative embodiment of a process for performing context switch sampling.
  • an addressable memory stores data and instructions for performing context switch sampling.
  • a processor includes hardware event counters for performance monitoring, and is coupled with the addressable memory to access said instructions and in response to said instructions, the processor counts occurrences of a first hardware event in a first hardware event counter and counts occurrences of a second hardware event in a second hardware event counter. After a specified number of occurrences of the first hardware event have been counted, it can be determined that a context switch has occurred. The second hardware event counter is then sampled and the hardware event counters are reset. In some embodiments the processor counts occurrences of segment register load events in the first hardware event counter and then records the sampled second hardware event counter value together with a process identifier value and/or a thread identifier value.
  • such techniques may be used to more accurately capture and measure events for a particular process and/or thread without including contamination from events captured from other processes and/or threads.
  • FIG. 1 illustrates one embodiment of a multiprocessing system 101 for performing context switch sampling.
  • Embodiments of system 101 may include addressable memory 140 having storage areas 141 - 149 to store data and machine executable instructions for performing context switch sampling.
  • Multiple processes 111 - 131 are coupled with addressable memory 140 via a bus or any other interconnect 110 and processes 111 - 131 concurrently execute their respective threads.
  • embodiments of processes 111 - 131 may include software processes or hardware threads or multiple individual processor cores on different dies or on the same die.
  • Some embodiments of system 101 may also include processors such as processor 102 , which has multiple hardware thread processes 111 and 121 .
  • Processes 111 - 131 respectively include execution units 112 - 132 , registers 113 - 133 , and hardware event counters 114 - 134 and 115 - 135 . It will be appreciated that in some embodiments one or more of execution units 112 - 132 may be physically shared by some of processes 111 - 131 . It will also be appreciated that in some embodiments of processes 111 - 131 , registers 113 - 133 , and/or hardware event counters 114 - 134 and 115 - 135 may also be shared or used in common by some of processes 111 - 131 .
  • One or more processes 111 - 131 may be coupled with the addressable memory 140 to access machine executable instructions 149 , and responsive to machine executable instructions 149 , one or more of processes 111 - 131 may count occurrences of a first hardware event in hardware event counters 114 - 134 respectively.
  • the first hardware event may be a segment register load event, as is used in context switches on “x86” processors, such as those manufactured by Intel Corp. of Santa Clara, Calif. It will be appreciated that by counting occurrences of a hardware event such as a segment register load event in hardware event counters 114 - 134 respectively, the one or more processes 111 - 131 may be able to determine when a context switch has occurred.
  • a given operating system running on a specific x86 processor may need to execute four segment register loads in order to perform a context switch between application processes at ring 3 .
  • SAV sample-after value
  • the value of the sampled hardware event counters more accurately capture and measure events for a particular process and/or thread without including contamination from events captured from other processes and/or threads.
  • a SAV count is set to eight (twice four) a context switch between processes will have been detected.
  • a SAV count is set to five (one more than four) a context switch between threads of the same process will have been detected.
  • a SAV count is set at or below the critical value (in this example, four) then sampling will fail to capture the desired application data, since sampling will occur prior to collecting event statistics in the desired application.
  • the number of segment register loads needed to detect a context switch may vary greatly (e.g. between 1 and over 100 segment register loads) from operating system to operating system depending on the particular operating system and on the particular processor.
  • a context switch may also be associated with loading of descriptor tables, and so triggering the sampling of event count data following the loading of descriptor tables (either global or local) may provide an alternative technique to perform context switch sampling.
  • a particular address in addressable memory 140 may be selected for the express purpose of triggering context switch sampling, whenever that address is accessed.
  • a hardware event trigger whenever a process and/or thread identifier is changed may be used, or a special instruction may be added to the processor architecture specifically to trigger context switch sampling.
  • the second hardware event counters, 115 - 135 respectively are sampled and the first and second hardware event counters 114 - 134 and 115 - 135 respectively, are reset.
  • embodiments of processes 111 and 121 may share or use in common hardware event counters 114 - 124 and 115 - 125 , so processor 102 , for example, may have just a single set of hardware event counters.
  • the sampled second hardware event counter 115 - 135 values may be recorded and/or accumulated in storage locations 143 - 147 with a process identifier 141 - 145 value in addressable memory 140 .
  • the sampled second hardware event counter 115 - 135 values may also be recorded and/or accumulated in storage locations 143 - 147 with a thread identifier 144 - 148 value in addressable memory 140 .
  • FIG. 2 illustrates a flow diagram for one embodiment of a process 201 to perform context switch sampling for performance monitoring in a multiprocessing system.
  • Process 201 and other processes herein disclosed are performed by processing blocks that may comprise dedicated hardware or software or firmware operation codes executable by general purpose machines or by special purpose machines or by a combination of both.
  • processing block 211 I occurrences of a first hardware event are counted in a first hardware event counter (e.g. one of ECs 114 - 134 ).
  • processing block 212 J occurrences of a second hardware event are counted in a second hardware event counter (e.g. one of ECs 115 - 135 ).
  • processing block 213 it is determined whether I is equal to a predetermined SAV value N. If not counting continues in processing block 211 . Otherwise processing proceeds to processing block 214 where the count J of second hardware event occurrences in the second hardware event counter is sampled.
  • processing block 215 the first and second hardware event counters are reset, and in processing block 216 , the sampled count J is recorded (e.g at locations 143 - 147 ) in addressable memory 140 with a process identifier (e.g 141 - 145 ) value.
  • FIG. 3 illustrates a flow diagram for an alternative embodiment of a process 301 for performing context switch sampling.
  • processing block 311 I occurrences of a segment register load event are counted in a first hardware event counter (e.g. one of ECs 114 - 134 ).
  • processing block 312 J occurrences of a second hardware event are counted in a second hardware event counter (e.g. one of ECs 115 - 135 ).
  • processing block 313 it is determined whether I is equal to a predetermined SAV value N. If not counting continues in processing block 311 . Otherwise processing proceeds to processing block 314 where the count J of second hardware event occurrences in the second hardware event counter is sampled.
  • processing block 315 the segment register load event count and the second hardware event count in the first and second hardware event counters are reset, and in processing block 316 , the sampled count J is recorded (e.g at locations 143 - 147 ) in addressable memory 140 with a process identifier (e.g 141 - 145 ) value.
  • processes 201 and 301 may be able to determine when a context switch has occurred, and hence be used to accurately capture and measure hardware events for a particular process and/or thread without including contamination from hardware events captured during execution of other processes and/or threads.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Quality & Reliability (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Computer Hardware Design (AREA)
  • Debugging And Monitoring (AREA)
US12/420,815 2009-04-08 2009-04-08 Context switch sampling Expired - Fee Related US8346509B2 (en)

Priority Applications (4)

Application Number Priority Date Filing Date Title
US12/420,815 US8346509B2 (en) 2009-04-08 2009-04-08 Context switch sampling
JP2010088905A JP5257816B2 (ja) 2009-04-08 2010-04-07 コンテキストスイッチサンプリング
EP10250736.5A EP2239664B1 (de) 2009-04-08 2010-04-08 Kontextwechselabtastung
CN201010207270.1A CN101859268B (zh) 2009-04-08 2010-04-08 上下文切换采样

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US12/420,815 US8346509B2 (en) 2009-04-08 2009-04-08 Context switch sampling

Publications (2)

Publication Number Publication Date
US20100262870A1 US20100262870A1 (en) 2010-10-14
US8346509B2 true US8346509B2 (en) 2013-01-01

Family

ID=42352696

Family Applications (1)

Application Number Title Priority Date Filing Date
US12/420,815 Expired - Fee Related US8346509B2 (en) 2009-04-08 2009-04-08 Context switch sampling

Country Status (4)

Country Link
US (1) US8346509B2 (de)
EP (1) EP2239664B1 (de)
JP (1) JP5257816B2 (de)
CN (1) CN101859268B (de)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8935675B1 (en) 2013-09-25 2015-01-13 Google Inc. Statistical sampling of event monitoring triggers under overhead constraints
US10680923B2 (en) 2016-02-11 2020-06-09 Samsung Electronics Co., Ltd. Semiconductor device and operating method thereof

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2544121B1 (de) 2010-03-03 2020-07-29 Panasonic Intellectual Property Management Co., Ltd. In einer aufzeichnungsmediumvorrichtung eingebettetes steuergerät, aufzeichnungsmediumvorrichtung, herstellungssystem für aufzeichnungsmediumvorrichtung und herstellungsverfahren für aufzeichnungsmediumvorrichtung
US9021172B2 (en) 2012-07-06 2015-04-28 Arm Limited Data processing apparatus and method and method for generating performance monitoring interrupt signal based on first event counter and second event counter
JP2014182478A (ja) 2013-03-18 2014-09-29 Fujitsu Ltd 性能プロファイリング装置及び性能プロファイリング方法
US9354883B2 (en) 2014-03-27 2016-05-31 International Business Machines Corporation Dynamic enablement of multithreading
US9218185B2 (en) 2014-03-27 2015-12-22 International Business Machines Corporation Multithreading capability information retrieval
US9921848B2 (en) 2014-03-27 2018-03-20 International Business Machines Corporation Address expansion and contraction in a multithreading computer system
US10102004B2 (en) 2014-03-27 2018-10-16 International Business Machines Corporation Hardware counters to track utilization in a multithreading computer system
US9804846B2 (en) 2014-03-27 2017-10-31 International Business Machines Corporation Thread context preservation in a multithreading computer system
US9417876B2 (en) 2014-03-27 2016-08-16 International Business Machines Corporation Thread context restoration in a multithreading computer system
US9594660B2 (en) 2014-03-27 2017-03-14 International Business Machines Corporation Multithreading computer system and program product for executing a query instruction for idle time accumulation among cores
US11200058B2 (en) * 2014-05-07 2021-12-14 Qualcomm Incorporated Dynamic load balancing of hardware threads in clustered processor cores using shared hardware resources, and related circuits, methods, and computer-readable media
CN105487958B (zh) * 2015-11-24 2018-04-10 无锡江南计算技术研究所 处理器内部行为监测方法
GB2550903B (en) * 2016-05-27 2019-06-12 Arm Ip Ltd Context data control
CN107451038B (zh) * 2016-05-30 2020-05-19 龙芯中科技术有限公司 硬件事件采集方法、处理器和计算系统

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05151004A (ja) 1991-11-30 1993-06-18 Nec Corp タスクのｃｐｕ占有時間測定方法
JPH07307727A (ja) 1994-05-13 1995-11-21 Nippon Motorola Ltd データ信号のサンプリング方法及びその回路
US5615135A (en) * 1995-06-01 1997-03-25 International Business Machines Corporation Event driven interface having a dynamically reconfigurable counter for monitoring a high speed data network according to changing traffic events
US6356615B1 (en) * 1999-10-13 2002-03-12 Transmeta Corporation Programmable event counter system
US6519310B2 (en) * 2001-03-28 2003-02-11 Intel Corporation Hardware event based flow control of counters
US20060167658A1 (en) 2005-01-27 2006-07-27 International Business Machines Corporation Method, apparatus, and computer program product in a performance monitor for sampling all performance events generated by a processor
US20080177756A1 (en) 2007-01-18 2008-07-24 Nicolai Kosche Method and Apparatus for Synthesizing Hardware Counters from Performance Sampling
US20090235056A1 (en) 2008-03-14 2009-09-17 Fujitsu Limited Recording medium storing performance monitoring program, performance monitoring method, and performance monitoring device
JP2010152458A (ja) 2008-12-24 2010-07-08 Fujitsu Ltd 性能測定プログラム及び性能測定方法並びに性能測定機能を有する情報処理装置。
US7779238B2 (en) * 2004-06-30 2010-08-17 Oracle America, Inc. Method and apparatus for precisely identifying effective addresses associated with hardware events

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1856612B1 (de) * 2005-01-28 2008-10-01 International Business Machines Corporation Zählverfahren für anweisungen zur protokollierung und wiedergabe einer deterministischen ereignisabfolge

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05151004A (ja) 1991-11-30 1993-06-18 Nec Corp タスクのｃｐｕ占有時間測定方法
JPH07307727A (ja) 1994-05-13 1995-11-21 Nippon Motorola Ltd データ信号のサンプリング方法及びその回路
US5615135A (en) * 1995-06-01 1997-03-25 International Business Machines Corporation Event driven interface having a dynamically reconfigurable counter for monitoring a high speed data network according to changing traffic events
US6356615B1 (en) * 1999-10-13 2002-03-12 Transmeta Corporation Programmable event counter system
US6519310B2 (en) * 2001-03-28 2003-02-11 Intel Corporation Hardware event based flow control of counters
US7779238B2 (en) * 2004-06-30 2010-08-17 Oracle America, Inc. Method and apparatus for precisely identifying effective addresses associated with hardware events
US20060167658A1 (en) 2005-01-27 2006-07-27 International Business Machines Corporation Method, apparatus, and computer program product in a performance monitor for sampling all performance events generated by a processor
US20080177756A1 (en) 2007-01-18 2008-07-24 Nicolai Kosche Method and Apparatus for Synthesizing Hardware Counters from Performance Sampling
US20090235056A1 (en) 2008-03-14 2009-09-17 Fujitsu Limited Recording medium storing performance monitoring program, performance monitoring method, and performance monitoring device
JP2009223451A (ja) 2008-03-14 2009-10-01 Fujitsu Ltd 性能モニタリングプログラム、性能モニタリング方法、性能モニタリング装置
JP2010152458A (ja) 2008-12-24 2010-07-08 Fujitsu Ltd 性能測定プログラム及び性能測定方法並びに性能測定機能を有する情報処理装置。

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
European Search Report for European Patent Application No. 10250736.5, mailed on Nov. 15, 2010, 4 pages.
Office Action Received for Chinese Patent Application No. 201010207270.1, mailed on Jan. 12, 2012, 7 pages of Office Action and 7 pages of English Translation.
Office Action Received for European Patent Application No. 10250736.5, mailed on Feb. 17, 2011, 6 pages.
Office Action Received for Japanese Patent Application No. 2010-088905, mailed on Nov. 1, 2011, 2 pages of Office Action and 2 pages of English Translation.
Office Action Received for Japanese Patent Application No. 2010-088905, mailed on Sep. 4, 2014, 2 pages of Office Action and 2 pages of English Translation.

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8935675B1 (en) 2013-09-25 2015-01-13 Google Inc. Statistical sampling of event monitoring triggers under overhead constraints
US10680923B2 (en) 2016-02-11 2020-06-09 Samsung Electronics Co., Ltd. Semiconductor device and operating method thereof
US11349738B2 (en) 2016-02-11 2022-05-31 Samsung Electronics Co., Ltd. Semiconductor device and operating method thereof
US11652718B2 (en) 2016-02-11 2023-05-16 Samsung Electronics Co., Ltd. Semiconductor device and operating method thereof
US12316520B2 (en) 2016-02-11 2025-05-27 Samsung Electronics Co., Ltd. Semiconductor device and operating method thereof

Also Published As

Publication number Publication date
EP2239664A2 (de) 2010-10-13
US20100262870A1 (en) 2010-10-14
CN101859268A (zh) 2010-10-13
EP2239664B1 (de) 2017-11-15
CN101859268B (zh) 2017-03-01
EP2239664A3 (de) 2010-12-15
JP5257816B2 (ja) 2013-08-07
JP2010244552A (ja) 2010-10-28

Similar Documents

Publication Publication Date Title
US20100262870A1 (en) Context switch sampling
US10140216B2 (en) Measuring address translation latency
US8566795B2 (en) Selectively obtaining call stack information based on criteria
CN101625657B (zh) 监控数据处理设备和概括监控数据
US9081629B2 (en) Excluding counts on software threads in a state
JP5299161B2 (ja) 計算機装置および消費電力のサンプリング方法
US9298651B2 (en) Continuous in-memory accumulation of hardware performance counter data
US20060005083A1 (en) Performance count tracing
US7315795B2 (en) Performance analyzing method using hardware
JP2020512613A5 (de)
US20090178036A1 (en) Method and Apparatus for Call Stack Sampling Using a Virtual Machine
US9600348B2 (en) Recording activity of software threads in a concurrent software environment
US9009537B2 (en) Diagnostic data capture in a computing environment
CN107451038B (zh) 硬件事件采集方法、处理器和计算系统
US6725363B1 (en) Method for filtering instructions to get more precise event counts
US9195524B1 (en) Hardware support for performance analysis
US20140075164A1 (en) Temporal locality aware instruction sampling
US8656411B2 (en) Technique for monitoring activity within an integrated circuit
US20090210196A1 (en) Method, system and computer program product for event-based sampling to monitor computer system performance
CN111858243A (zh) 基于指数增长的多硬件事件监控计数值估计方法
CN115080577B (zh) 数据统计周期切换方法、装置、设备及存储介质
US10657021B2 (en) Tracing processing activity with position data to reorder items of trace data
CN112817826A (zh) 一种数据整合方法及装置

Legal Events

Date Code Title Description
AS Assignment

Owner name: INTEL CORPORATION, CALIFORNIA

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:DAVIES, ROBERT;REEL/FRAME:024051/0972

Effective date: 20090505

FEPP Fee payment procedure

Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

STCF Information on status: patent grant

Free format text: PATENTED CASE

FPAY Fee payment

Year of fee payment: 4

MAFP Maintenance fee payment

Free format text: PAYMENT OF MAINTENANCE FEE, 8TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1552); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

Year of fee payment: 8

FEPP Fee payment procedure

Free format text: MAINTENANCE FEE REMINDER MAILED (ORIGINAL EVENT CODE: REM.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

LAPS Lapse for failure to pay maintenance fees

Free format text: PATENT EXPIRED FOR FAILURE TO PAY MAINTENANCE FEES (ORIGINAL EVENT CODE: EXP.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

STCH Information on status: patent discontinuation

Free format text: PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362

FP Lapsed due to failure to pay maintenance fee

Effective date: 20250101